Math 211 Fall 2007: Solutions: HW #3 Instructor: S. Cautis 1. section 2.6, #2 dF = (2x − y)dx + (−x + 2y)dy 2. section 2.6, #16 2u (2u)(2v) ∂ 2v ∂ ( )= 2 = ( ) ∂v u2 + v 2 (u + v 2 ) ∂u u2 + v 2 so the equation is exact! Solving ∂ ∂u F = 2u u2 +v 2 by integrating with respect to u we get F (u, v) = ln(u2 + v 2 ) + φ(v) Taking the derivative with respect to v we must get 2v 2v + φ0 (v) = 2 u2 + v 2 u + v2 so φ(v) = C is a constant. Thus F = ln(u2 + v 2 ) + C. 3. section 2.6, #24 Integrating 3(y+1)2 x4 with respect to x we get (y + 1)2 + φ(y) x3 Taking the derivative with respect to y we get F (x, y) = − − 2(y + 1) 2x(y + 1) + φ0 (y) = − x3 x4 2 from which we get φ0 (y) = 0 so φ(y) = C constant. Thus F = − (y+1) x3 +C. 4. section 2.6, #28 We want µ(y) such that ∂ ∂y (2yµ) = ∂ ∂x ((x + y)µ) or equivalently 2µ + 2yµ0 = µ √ Thus 2yµ0 = −µ. Solving by separating variables we get µ = 1/ y. Thus we need to solve √ √ 2 ydx + (x + y)/ ydy = 0 Integrating the first term with respect to x we get √ F = 2x y + φ(y) and taking the derivative with respect to y gives √ √ x/ y + φ0 = (x + y)/ y √ or φ0 = y which means φ = 2/3y 3/2 + C. 5. section 2.6, #36 A quick check shows it is not exact. Fortunately, it is homogeneous so we can make the substitution y = xv to get (x + xv)dx + (xv − x)(xdv + vdx) = 0 or equivalently (1 + v)dx + (v − 1)(xdv + vdx) = 0 or (1 + v 2 )dx + x(v − 1)dv = 0 which is separable. We separate to get dx/x = −(v − 1)/(1 + v 2 )dv so integrating we get ln |x| = −1/2 ln(1 + v 2 ) + arctan(v) + C Substituting back v = y/x we get the general solution − ln |x| − 1/2 ln(1 + (y/x)2 ) + arctan(y/x) + C = 0 6. section 2.7, #1 f (t, y) = 4 + y 2 is continuous everywhere and so is unique solution is guaranteed. ∂ ∂y f (t, y) = 2y so a 7. section 2.7, #2 √ ∂ 1 f (t, y) = y is continuous everywhere and ∂y f (t, y) = 2√ y is continous and well defined for y > 0. Our initial condition is at t0 = 4 and y0 = 0 and so there is no rectangle containing this point where both are continuous – so hypothesis of theorem is not satisfied. 8. section 2.7 #6 ∂ f (x, y) = −y/x2 is also f (x, y) = y/x + 2 is continuous for x 6= 0 while ∂x continuous for x 6= 0. Since initial condition is (x0 , y0 ) = (0, 1) there is no rectangle again and so theorem does not guarantee a unique solution. 9. section 2.7 #8 The equation is linear. The general solution is y(t) = t + 2Ct2 . All of these solutions pass through the point (0, 0) and look like (are in fact) a parabola. This means that y(0) = 2 has no solution. If we put this equation into normal form we get dy = 2y/t − 1 dt which is not continuous at t = 0. Consequently the hypothesis of the existence theorem are not satisfied. 10. section 2.7 #21 (a) We have y 0 (t+ 0) = = lim+ t→t0 (t − t0 )3 − 0 t − t0 lim (t − t0 )2 t→t+ 0 = 0 Similarly, y 0 (t− 0) = lim t→t− 0 = 0 0−0 t − t0 So the two limits agree and thus y(t) is differentiable at t = t0 with y 0 (t0 ) = 0 since both right and left hand derivatives are equal to 0. (b) In this case f (t, y) = 3y 2/3 is continuous at (0, 0) (so solution exists ∂ by theorem) but ∂y f (t, y) = 2y −1/3 which is not continuous at (0, 0). Hence theorem 7.16 does not guarantee a unique solution there since the hypothesis is not satisfied (and indeed, as we see, there are at least two solutions).